Privacy Policy
Last updated: 27 September 2026
seedoes is a footage-search and storyboarding tool for directors and production teams, operated by GoGo LTD. (甲又夠夠有限公司), a company registered in Taiwan. This policy explains what we collect, why, who we share it with, and how you can get it removed.
- With the macOS app, your original video files never leave your computer. Only still frames are uploaded, plus, for each shot, a short low-resolution excerpt of about two seconds that is used once to describe movement and then deleted.
- In a web browser, you upload the whole video file. We extract frames from it and then delete the file — it is not kept once processing has finished.
1. Who we are
GoGo LTD. is the data controller for the information described here. You can reach us at support@seedoes.ai. Our full registration details are at the bottom of this page.
2. What we collect
2.1 Account information
- Your email address, used to identify your account and to contact you.
- A display name you choose.
- Your job title and the type of company you work in, chosen from a list when you sign up (with a free-text field if you pick "Other"). We use them to understand who uses seedoes, and we may publish aggregate figures such as the share of users who are directors; we never identify you. Both are required to start using the service. You can change them later in the app's settings.
- Interface and tagging language preferences.
- Your usage allowances and points, the version of the seedoes application you are running, the dates your account was created and last used, and which version of these documents you agreed to and when.
- Account events such as sign-ins, an upload refused for lack of points, or a file that keeps failing, kept so that we can support you and spot problems.
Sign-in is handled by Supabase. We do not receive or store your password. If you sign in with Google, we receive the email address and name on your Google account, nothing else.
2.2 Material you add to your library
Whichever way you add footage, we end up holding the same things (the excerpts below are the one exception: they pass through and are not kept):
- Still frames (JPEG images) selected from the video.
- Smaller copies of those frames that we make so that lists of thumbnails load faster.
- Short excerpts (MP4) — with the app, one per shot, about two seconds long, reduced to 480 pixels high. They are used once, to describe movement, and deleted as soon as the description is done; we do not keep them. Browser uploads produce still frames only, no excerpts.
- File information: filename, duration, resolution, frame rate, file size, and a checksum of the file contents. The checksum lets us recognise the same file after you rename or move it, so your labels are not lost and the work is not repeated.
What differs is whether the original file is uploaded at all:
- seedoes app for macOS. The application analyses each video on your own computer and uploads only the frames, excerpts and file information above. Your original files are never uploaded. Playback always plays the file on your own disk.
- seedoes in a web browser. The whole video file is uploaded to our storage so that we can extract frames from it. The file is deleted as soon as processing finishes, whether it succeeded or failed. We do not keep a copy, and the file is not used for anything other than extracting the frames.
2.3 Descriptions generated from your material
Each uploaded frame is described along nine dimensions — scene, people, clothing, action, on-screen text, style, emotion, colour and camera — together with a short summary of the video. These descriptions, and the numerical representations (embeddings) derived from them, are what make search work.
2.4 How you use the service
- Your search queries, the language detected, the mode used (still or motion), how many results were returned and how long the search took.
- Which result you opened, and its position in the list. We use this to measure and improve search quality.
- Records of calls made to AI providers on your behalf — provider, model, token counts and cost — so that we can monitor and control operating costs.
- Storyboard briefs you submit and the storyboards produced from them.
- The days on which you used the service, and the country or region you connected from — a two-letter country code determined by Cloudflare when it forwards the request. We use it to understand where the service is being used.
2.5 What we do not collect
- We do not keep your original video files, unless you are on a plan that explicitly includes cloud storage of originals. No current plan does. If that changes, we will tell you before it applies to your account. With the app your originals are never uploaded; with the browser they are deleted once frames have been extracted.
- We do not collect the folder paths or directory structure of your disk. Folder locations are recorded only on your own computer.
- We do not use tracking or advertising cookies. A sign-in token is stored locally on your computer so that you stay logged in.
- We do not record or store your IP address. The country code is added by Cloudflare when it forwards the request; those two letters are all we receive.
3. Who processes your data
We use the following providers to run the service. Frames (and, with the app, the short excerpts) from your footage are sent to the AI providers marked below so that they can be described.
| Provider | Purpose | Receives your material? |
|---|---|---|
| Supabase | Accounts and sign-in | No |
| Google (Sign in with Google) | Optional sign-in | No |
| Cloudflare (R2) | Stores frames and their smaller copies; briefly holds an uploaded video or frame package until processing is done | Yes |
| Railway | Application hosting and database (Singapore) | Yes |
| OpenRouter | Routes requests to the AI models below | Yes |
| Google (Gemini) | Describes frames and excerpts | Yes |
| Anthropic (Claude) | Generates storyboards | Briefs only |
| Jina AI | Converts text and images into search vectors | Yes |
| Paddle | Payment processing | No |
These providers operate outside Taiwan: our servers and database run in Singapore, and the others operate in places including the United States and the European Union. Using seedoes therefore involves transferring your data internationally. Each provider handles data under its own terms, and we recommend reading them if this matters to you.
We do not sell your data, and we do not use your material to train our own models. We cannot make commitments on behalf of the providers listed above; their own terms govern what they do with data they receive through their APIs.
4. Why we are allowed to process it
- To provide the service you asked for. Storing and describing frames is the service; without it there is nothing to search.
- Our legitimate interest in keeping the service working, controlling operating costs, and improving search quality.
- Legal obligations, such as keeping payment records.
5. How long we keep it
- Account information and library data are kept for as long as your account exists.
- When you delete your account, or ask us to delete it, the frames stored for you, together with their descriptions, are removed within 30 days. The delay is a safety margin against accidental deletion.
- Our database is backed up every day. A backup can hold account information and descriptions for up to 35 days after they were removed from the live system; it is used only to restore the service after a failure. Frames and their smaller copies are not backed up: once removed, they are gone.
- Anonymised operational records — how long searches took, how much a call cost — may be kept after that for capacity and cost planning. These contain no material from your footage.
- Payment records are kept for as long as tax law requires.
6. Your rights
You can ask us to show you what we hold about you, correct it, delete it, or provide it in a portable form. You can also object to processing based on our legitimate interests. Write to support@seedoes.ai and we will respond within 30 days. If you are in the EU or UK and are not satisfied with our response, you may complain to your local data protection authority.
7. Material belonging to other people
Footage held by a director or production company usually belongs to a client. Before you add material to seedoes, make sure you are permitted to have it described by the third-party providers listed in section 3. If your agreement with a client forbids sending their material to external services, do not add that material to your library.
8. Security
Frames and their smaller copies are kept in private storage with no public address. They can be reached only through links that our server issues after you sign in, for your own material, and each link expires after ten minutes. Storage is encrypted at rest (AES-256) and all traffic is encrypted in transit (TLS). Access to production systems is limited to GoGo LTD. personnel who need it. No system is perfectly secure; if a breach affects your data, we will tell you and the relevant authority as required by law.
9. Children
seedoes is a professional tool and is not intended for anyone under 18. We do not knowingly collect information from children.
10. Changes
If we change this policy in a way that materially affects you, we will email the address on your account before the change takes effect. The date at the top of this page always shows the current version.
11. Contact
隱私權政策
最後更新：2026 年 9 月 27 日
seedoes 是給導演與製作團隊用的素材搜尋與分鏡工具，由在台灣登記的 甲又夠夠有限公司（GoGo LTD.）營運。這份文件說明我們收集什麼、 為什麼收、交給誰處理，以及你要怎麼把它刪掉。
- 用 macOS app：你的原始影片檔案完全不會離開你的電腦。 上傳的只有靜態影格，以及每個鏡頭約兩秒的低解析度短片段—— 那段片段只用來描述動態一次，描述完就刪除。
- 用瀏覽器：整支影片會上傳。我們從裡面抽出影格之後 就把檔案刪掉——處理完就不留了。
1. 我們是誰
甲又夠夠有限公司是本文所述資料的控制者。聯絡方式： support@seedoes.ai。 完整登記資料在本頁最下方。
2. 我們收集什麼
2.1 帳號資料
- 你的 email，用來識別帳號與聯絡你。
- 你自己設定的顯示名稱。
- 你的職稱與公司類別，註冊時從清單中選（選「其他」可以自己填）。 我們用它了解 seedoes 是誰在用，也可能公布整體統計，例如「使用者中有四成是導演」， 但不會指出你個人。這兩項是開始使用前必填的；之後在 app 的設定頁可以改。
- 介面語言與標籤語言的偏好。
- 你的使用額度與點數、你在跑的 seedoes 版本、帳號建立與最後使用的時間， 以及你同意了哪一版文件、什麼時候同意的。
- 帳號事件，例如登入、因點數不足被擋下的上傳、一直失敗的檔案——用來協助你、也用來發現問題。
登入由 Supabase 處理，我們不會拿到、也不會保存你的密碼。 用 Google 登入的話，我們會從 Google 取得你的 email 與名稱，其他都不會。
2.2 你加進素材庫的東西
不論用哪一種方式，我們最後保存的都是同一批東西（短片段是例外：只經過、不保存）：
- 靜態影格（JPEG 圖片），從影片裡挑出來的關鍵畫面。
- 縮小的副本：我們從影格另外做出來的小圖，讓縮圖列表載得快。
- 短片段（MP4）——用 app 時，每個鏡頭一段、約兩秒、縮到 480 像素高。 只用來描述動態一次，描述完就刪除，我們不保留。瀏覽器上傳只留影格，沒有短片段。
- 檔案資訊：檔名、長度、解析度、影格率、檔案大小，以及檔案內容的 指紋。有這個指紋，你改檔名或搬資料夾之後我們才認得出是同一支， 標籤不會不見、也不用重做一次。
差別在於「原始檔案本身會不會被上傳」：
- seedoes macOS app：程式在你自己的電腦上分析每一支影片， 只上傳上面那些影格、短片段與檔案資訊。你的原始檔案完全不會被上傳。 播放時播的一律是你硬碟上的檔案。
- 用瀏覽器使用 seedoes：整支影片會上傳到我們的儲存空間， 我們才能從裡面抽影格。處理一結束就立刻刪掉那個檔案， 不論成功或失敗。我們不留副本，那個檔案也不會被用在抽影格以外的任何用途。
2.3 從素材產生的描述
每一張上傳的影格會被描述成九個面向——場景、人物、服裝、動作、畫面文字、風格、 情緒、色彩、鏡頭——再加上一段影片摘要。這些描述以及由它們算出來的數值向量， 就是搜尋能運作的原因。
2.4 你怎麼使用這個服務
- 你打的搜尋字串、偵測到的語言、用的模式（STILL 或 MOTION）、回傳幾筆、花了多久。
- 你點開了哪一筆、它排在第幾位。我們用這個衡量與改善搜尋品質。
- 代你呼叫 AI 供應商的紀錄——供應商、模型、token 數、成本——用來控制營運費用。
- 你送出的分鏡 brief 與產生的分鏡結果。
- 你在哪一天使用了服務，以及連線來源的國家或地區——兩個字母的國碼， 由 Cloudflare 在轉送請求時判定。我們用它了解服務在哪些地區被使用。
2.5 我們不收集的
- 不保存你的原始影片檔案，除非你使用的方案明確包含「原始影片雲端保存」—— 目前沒有任何方案包含。若之後有所改變，我們會在它適用到你的帳號之前先告訴你。 用 app 是根本不會上傳；用瀏覽器是抽完影格就刪掉。
- 不收集你硬碟上的資料夾路徑與目錄結構。資料夾位置只記在你自己的電腦裡。
- 不使用追蹤或廣告 cookie。登入憑證存在你自己的電腦上，讓你不用每次重登。
- 不記錄、也不保存你的 IP 位址。國別是 Cloudflare 轉送請求時附上的兩碼國碼， 我們只收到那兩個字母。
3. 誰會處理你的資料
我們用下列供應商來營運這個服務。表格中標示「是」的供應商，會收到從你素材裡 抽出來的影格（用 app 時還有短片段），因為要由他們來描述。
| 供應商 | 用途 | 會收到你的素材嗎 |
|---|---|---|
| Supabase | 帳號與登入 | 否 |
| Google（Google 登入） | 可選的登入方式 | 否 |
| Cloudflare（R2） | 存放影格與縮小的副本；上傳的影片或影格包會在這裡 短暫停留到處理完為止 | 是 |
| Railway | 應用主機與資料庫（新加坡） | 是 |
| OpenRouter | 把請求轉給下面的 AI 模型 | 是 |
| Google（Gemini） | 描述影格與短片段 | 是 |
| Anthropic（Claude） | 生成分鏡 | 只有 brief |
| Jina AI | 把文字與影像轉成搜尋向量 | 是 |
| Paddle | 金流處理 | 否 |
這些供應商都在台灣境外營運：我們的主機與資料庫在新加坡，其他供應商在美國、歐盟等地。 使用 seedoes 因此會發生資料的跨境傳輸。每一家都依照自己的條款處理資料，如果你在意這件事，建議去看一下他們的條款。
我們不販售你的資料，也不會拿你的素材訓練我們自己的模型。 我們無法代上述供應商做出承諾——他們透過 API 收到的資料如何處理，依他們自己的條款。
4. 我們憑什麼可以處理
- 提供你要求的服務。存放與描述影格就是這個服務本身， 沒有它就沒有東西可以搜。
- 正當利益：維持服務運作、控制營運成本、改善搜尋品質。
- 法律義務，例如保存交易紀錄。
5. 保存多久
- 帳號資料與素材庫資料，在你的帳號存在期間持續保存。
- 你刪除帳號、或要求我們刪除之後，為你存放的影格與相關描述會在 30 天內清除。這段緩衝是為了防止誤刪。
- 資料庫每天備份一次。備份裡的帳號資料與描述，在正式系統清除之後最多再保留 35 天， 只用於故障後的還原。影格與縮小的副本沒有備份：清除了就沒有了。
- 去識別化的營運紀錄（搜尋花了多久、一次呼叫多少錢）可能會在那之後繼續保留， 用於容量與成本規劃。這些紀錄不含你素材的任何內容。
- 交易紀錄依稅法規定的期間保存。
6. 你的權利
你可以要求我們告訴你保存了哪些關於你的資料、更正它、刪除它，或以可攜格式提供給你， 也可以對基於正當利益的處理表示反對。寫信到 support@seedoes.ai，我們會在 30 天內回覆。 若你在歐盟或英國，且對我們的回覆不滿意，可以向你當地的資料保護主管機關申訴。
7. 屬於別人的素材
導演或製作公司手上的素材，通常所有權在客戶那邊。把素材加進 seedoes 之前， 請先確認你有權把它交給第 3 節列出的第三方供應商描述。如果你跟客戶的合約禁止 把素材送到外部服務，就不要把那批素材加進素材庫。
8. 安全
影格與縮小的副本放在沒有公開網址的私有儲存空間。只有你登入後，由我們的伺服器針對你自己的素材 簽發的連結才拿得到，而且每一條連結十分鐘就失效。儲存時一律加密（AES-256），傳輸全程加密（TLS）。 正式環境的存取權限，僅限甲又夠夠有限公司內部有需要的人員。沒有任何系統是絕對安全的；若發生外洩且影響到 你的資料，我們會依法通知你與主管機關。
9. 未成年人
seedoes 是專業工具，不提供給未滿 18 歲者使用。我們不會刻意收集兒童的資料。
10. 變更
若本政策的變更對你有實質影響，我們會在生效前寄信到你帳號上的 email。 本頁最上方的日期永遠代表現行版本。
11. 聯絡
- Company / 公司
- GoGo LTD.（甲又夠夠有限公司）
- Address / 地址
- 2F., No. 116, Taiping Rd., Shulin Dist., New Taipei City 238016, Taiwan (R.O.C.)
- Reg. no. / 統一編號
- 89159100
- Contact / 聯絡
- support@seedoes.ai